Effect of load transfer by friction on the fatigue behaviour of riveted lap joints

Abstract An experimental investigation of load transmission throughout a riveted lap joint is presented with a focus on the role of friction. The main objective was to generate results to be utilized in a semi-empirical model for predicting the fatigue life of riveted joints, which is being developed by the authors. The test variables accounted for in the experiments were the rivet squeeze force and the faying surface condition. Triple-row riveted lap joint coupons of aluminium alloy 2024-T3 Alclad sheets intended to be a simple representation of longitudinal connections in the aircraft fuselage skin structure were assembled with two different rivet squeeze forces. For either force half of the joints had a thin Teflon interfoil in the overlap area in order to minimize friction between the sheets. Friction measurements reveal that, contrary to the friction coefficient between the Alclad mating surfaces, the friction coefficient between the Teflon and Alclad contact surfaces is considerably lower and does not increase with the cycle number. A convenient method of the determination of load transmission throughout a lap joint is proposed utilizing measurements by strain gauges mounted on only outer surfaces of the overlapping sheets supplemented by the secondary bending analysis according to a simple theoretical model. The results obtained indicate that the effect of the squeeze force and the faying surface condition on axial forces in the overlapping sheets is quantitatively insignificant. It is noted that an analytical procedure commonly used to compute axial forces in the sheets of joints with mechanical fasteners considerably underestimates load transfer by the critical, outer rivet rows. Comparative fatigue tests on the riveted coupons with the Teflon interlayer and without such a layer reveal superior fatigue lives of the former. These results demonstrate, contrary to the prevailing opinion, an overall detrimental influence of frictional load transfer at applied stress values representative of those experienced in service by lap splices of the aircraft fuselage. Post-failure examinations of the riveted joints provide evidence that the crack initiation locations and crack paths depend on the test variables and the stress level. For the uncoated joints, the effect of the squeeze force and applied load on the area of fretted regions around the rivet holes was noted, whilst the total roughness heights did not exhibit such a dependency. It is shown that the present and reported in other works experimental results on the effect of coatings on the fatigue performance of riveted joints cannot be explained based on FE stress analyses available in the literature due to simplifications involved in modelling the faying surface condition. A major conclusion of the present work is that because of the extremely complex determinants of frictional load transfer, its effect on the fatigue behaviour of riveted joints can only be predicted using a semi-empirical approach.